READINESS AUDIT · $499 · 72 HOURS

Know if your app is ready before you launch, raise or sell.

You built it fast with an AI app builder. Before real users, an investor or a buyer looks closely, get a plain-English verdict: ship, fix first, or do not ship. With the top risks explained and a fix list in the order to do it.

Written by Dark Vector Cognition LLC, Texas. Every finding points to where it lives, the file and line or the database rule, so you or your developer can check it.

THE OFFER

One app. One verdict. Three days.

One app, one verdict, in three days, before the moment that makes the risks expensive.

FOUNDING PRICE

$499 per app.

Founding price $499 for the first 5 audits; $999 after.

A written audit of one app, ending in a verdict you can act on and a fix list your developer, or your AI builder, can work through.

  • SCOPE / One app: one repository, or one front end and its backend
  • INPUT / A repo link or app URL, plus access to the code
  • OUTPUT / A written report by email: verdict, top risks, fix list
  • TIME / Within 72 hours of access to the code, or your money back

WHAT WE LOOK FOR

The six ways these apps most often go wrong.

Public security research keeps finding these in apps made with AI app builders. In one 2026 study of apps on Supabase, a database many of these builders use, 16,326 databases exposed tables anyone could read. The audit checks each of the six against your code.

01

Database rules left open

Tables that anyone with the app's public key can read or change, or rules that exist but do not limit anything.

02

Secret keys in the app

Admin, payment or AI provider keys shipped in the code that every visitor downloads, or left in the project history.

03

Server routes that skip sign-in

Functions and API routes that trust a user id sent by the browser instead of checking who is asking.

04

Public file storage

Uploads such as documents, photos or invoices that anyone with the link can open.

05

Payments that do not deliver

Checkouts that take money with no reliable record of who paid, and prices or quantities the browser can change.

06

Promises the code does not keep

Screens, docs and success pages that say more than the app actually does, including placeholder numbers shown as real data.

WHAT YOU GET

A report you can read without a security background.

Every finding says what is wrong, what it means for you, and how to check that the fix worked.

THE VERDICT

Ship, fix first, or do not ship, with the one reason that decides it and what would change it.

THE TOP RISKS

Up to five, each in plain words: what is wrong, what it means for you, where it is in the code, how to fix it, and how to check the fix.

THE FIX LIST

Every finding in the order to fix it, with the effort for each and whether it changes the verdict.

WHAT IS SOUND

What we checked and found in good order, so you know what the audit covered.

WHAT WE COULD NOT CHECK

Anything outside the code you shared, named so nothing is assumed.

A SHORT GLOSSARY

Every technical term in the report, explained with an everyday example.

ILLUSTRATION · NOT A REAL APP

VERDICT: DO NOT SHIP

1. Anyone can read your customers' orders (Critical)

  • WHAT IS WRONG / The orders table has row rules switched on, but its read rule allows every signed-in user, so one customer can list every other customer's orders.
  • WHAT IT MEANS FOR YOU / Any customer who signs up can see names, addresses and order history for all of your customers.
  • WHERE / supabase/migrations/0003_orders.sql, line 14
  • HOW TO FIX IT / Change the read rule to auth.uid() = user_id, so each customer sees only their own orders.
  • HOW TO CHECK THE FIX / Sign in as customer B and open customer A's order page. You should see nothing.

WHAT WE NEED FROM YOU

A link. A way to read the code if it is private.

The code gives a full read of the app. If your app uses Supabase, your database rules complete it.

REQUIRED

Your repo link or app URL

A link to the repository. For a private repo, the page after checkout and the confirmation email show how to share it. If you have no repo, an export from your builder works. You paste the link at checkout, with what you built it with and what you plan to do next.

WE ONLY READ
IF YOU USE SUPABASE

Your database rules

The page after checkout and the confirmation email include two short read-only queries that list your tables, their rules and your storage buckets. Paste its result into your reply. Rules changed in the dashboard do not show in the code, so without this export the best verdict we can give is fix first.

WE ONLY READ

WHO IT IS FOR

The founder with a deadline and an app nobody has checked.

Founders and small teams who built an app with Lovable, Bolt, v0, Replit or Cursor and are about to launch it, raise money on it, or sell it.

IF YOU ARE LAUNCHING

What a user, or someone poking at your app, could do on day one, and what to fix before you open sign-ups.

IF YOU ARE RAISING MONEY

What an investor's technical diligence will find, and the answers to have ready.

IF YOU ARE SELLING THE APP

What a buyer's diligence will find, and how each finding can move the price.

FOR TEAMS · DEEP PASS · WAITLIST

Keep your review bot. Add a deep pass on the pull requests that can hurt you.

Review bots are good at breadth: many small, real issues on every pull request. The defects that cause incidents are often deeper: data that ends up in the wrong place, two requests that race, a check that one path skips.

$2,000 PER TEAM PER MONTH

A second, deeper review, only where it matters.

Deep Pass is a GitHub App. It runs only on pull requests that touch user data, files, sign-in and permissions, work that runs at the same time or money, and posts its findings as pull request comments next to your current reviewer. It does not replace that reviewer.

Deep Pass is not open yet. Join the waitlist and we will write to you when it is.

Please leave out code, passwords and keys.

We use your details only to tell you when Deep Pass opens. Privacy policy.

TERMS, IN PLAIN WORDS

What you are agreeing to.

Short, because the offer is simple.

  • One audit covers one app: one repository, or one front end and its backend.
  • We read the code and configuration you share. We never send requests to your live app, test attacks on it, or change anything.
  • If your app is too large to audit well in 72 hours, we tell you within 24 hours of receiving access and refund you in full.
  • The 72 hours start when we can read your code. If the report is late, you get your money back.
  • If we cannot read your code within 7 days of your order, you get a full refund.
  • Your code is used only to write your report. We do not sell it, and we share it only with the service providers we use to do the work. We delete our copy 30 days after delivery, or sooner if you ask.
  • The audit is not a penetration test, and it is not legal, compliance or financial advice. No audit can prove that an app has no defects.
  • Our liability is limited to the fee you paid.
  • Questions before or after you order: hello@darkvectorcognition.ai

$499 · ONE APP · 72 HOURS OR YOUR MONEY BACK

Order an audit.

Founding price $499 for the first 5 audits; $999 after.

Paste your repo link or app URL at checkout and pay. The report arrives by email within 72 hours of access to your code, or you get your money back.